Reliability Sensitivity Estimation of Complex Models: A Simulation-Based Approach

One of the objectives of reliability sensitivity analysis is to study the influence of probabilistic model parameters onto the reliability of a given structural system. In this context, the system parameters involved in the sensitivity analysis are modeled by a random vector whose joint probability distribution is explicitly known and dependents on a certain number of parameters. The effect of distribution parameters on the system reliability is obtained by calculating the partial derivative of the failure probability with respect to such parameters. Of practical importance is the reliability sensitivity analysis of medium/large nonlinear finite element models subject to stochastic excitation. This type of problems appears in a number of realistic situations related to for example, earthquake engineering, offshore engineering, wind engineering, etc. The determination of the variation in the reliability (or equivalently in the failure probability) due to changes in system parameters can provide useful information. For example it can be used to identify the most influential system parameters and it can provide an important insight on system failure for risk-based decision making, such as robust control or reliability-based design optimization.